91
92
conductivity and therefore raised the limits of detection and quantification to 1-16 µM and 4-50 µM, respectively.
It has been shown that by changing the composition of the mobile phase, salt or acid concentration and content of organic modifier, this method can be universally applied to IL ions independent of the presence of UV-active elements or limitations to the hydrophobicity.
When injected as single compounds, the polar ions are favored due to lower acetonitrile content. However, the more hydrophobic ions are easier to separate from inorganic matrices, whereas the hydrophilic ones are challenging in this specific task. Here, hydrophilic interaction liquid chromatography (HILIC) has already proven to be an ideal alternative.29,132 Moreover, compared to IC, HPLC is preferred by higher throughput, smaller sample volume and lower prices for column and sample-preparation material, e.g. filter and pre-column. On the other hand, IC can be equipped with modules for inline sample preparation like filtration, dialysis, matrix elimination, preconcentration or heavy metal removal. This could be advantageous for specific industrial applications where such pretreatment is essential and would reduce time and costs. The usage of a suppressor for the analysis of anions results in a salt free mobile phase and would allow the linking to a MS detector. This enables an efficient and reliable identification of the analytes if unknown components are present (e.g.
degradation products or contaminants).
2. Cyano- and fluoro-based IL anions The cyano-based IL anions N(CN)2
-, C(CN)3
and B(CN)4
-, as well as the fluorinated anions (CF3SO2)2N-, (C2F5)3PF3
and H(C2F4)SO3
-, are frequently mentioned in the literature as suitable counter ions to design ILs with excellent electrochemical properties to be used in dye-sensitized solar cells or batteries.166 However, only little was known for their ecotoxicity (apart from (CF3SO2)2N-) and stability. Paper No. 7 summarized the results from biodegradation experiments and it has to be concluded that none of the anions are biodegradable under aerobic or nitrify conditions. Furthermore, the hydrolysis studies in Paper No. 2 showed the stability of these anions in environmental media (pH value of 7 and 9). On the other hand, harsh conditions (pH 1 and 13), as could maybe be found in technical applications, led to complete hydrolysis of N(CN)2
and C(CN)3
-, whereas all other anions can be assumed to be stable at ambient temperatures for at least one year. Kinetic studies for the
93
hydrolysis rates at different temperatures have been performed for aforementioned unstable anions. By MS measurements, some hydrolysis products could be identified and a degradation mechanism was predicted. Theoretical calculations of pKa values support this pathway and could help to explain observed contrary hydrolysis velocities for both anions in the investigated pH environments. The ecotoxicity studies summarized in paper No. 3 showed an evident acute effect for (C2F5)3PF3
in all aquatic test systems (V. fischeri, S.
vacuolatus, D. magna and L. minor), whereas N(CN)2
was similarly toxic compared to the IL containing chloride as counter ion. The other three anions had various influences, highly depending on the test organism.
Until now, ILs have not been determined in the environment and, unless they are used only in small scale, their entry in large amount is unexpected. Though, raising commercialization would require an assured waste management in order to minimize the risk of being discharged into the environment (e.g. by using the compound in closed systems or recycling them). Within REACH, an exposure analysis is required for the registration of chemicals produced in amounts of > 1 t a-1. This is of especially high concern for (C2F5)PF3- and (CF3SO2)2N- since they showed severe ecotoxicities in aquatic test systems. Furthermore, highly halogenated compounds (e.g. perfluorooctane sulfonic acid, polychlorinated biphenyls or dichlordiphenyltrichlorethan) are already known to be hardly degradable in the environment and therefore restricted or eliminated by the Stockholm Convention.167 Also, the results within this thesis illustrated that, most likely, none of these anions is degraded when they are released into the environment. This might lead to persistence, enrichment and/or, for the more hydrophobic anions, to bioaccumulation. However, for such statements, further testing is necessary. Investigations with e.g. inoculum pre-adopted to the test substance or biodegradation under realistic conditions in a waste water treatment plant are needed for a sound hazard assessment. Furthermore, alternative degradation pathways, like UV-treatment and/or advanced oxidation processes (AOP), should be considered.
Considering toxicity testing, long term effects at sub-acute concentrations or mixture effects of these anions among each other or with other environmental chemicals, still need to be investigated.
The results of the ecotoxicity studies presented in paper No. 3 indicate a lower ecotoxicological hazard potential for the cyano based anions N(CN)2
-, C(CN)3
and B(CN4)-.
94
However, the technical application of N(CN)2
and C(CN)3
in strongly acidic or basic conditions might pose problems through hydrolysis. Even though the hydrolysed solutions showed similar or less cytotoxicity compared to the non-hydrolysed one, adverse effects cannot be excluded since some of the identified/predicted hydrolysis products are classified as toxic. Additional, the reduced lifetime might lead to loss in performance or machine damage and, thus to more waste and costs. The hydrolysis and biodegradation studies presented here hypothesize a persistence of these anions in the environment. Recently it was shown that enzymatic hydrolysis by nitrile hydratase led to the corresponding amides.168 This indicates degradation in the environment when microorganisms containing this enzyme are present or opens the possibility for effluent treatment.
Slight modifications in the chemical structure of these anions are an option to design alternatives. The partial replacement of the perfluorinated chains with alkyl chains or the introduction of functional groups could be a possibility to design structures with reduced hazard. Hydroxyl groups, for instance, can serve as a contact point for biological breakdown and will reduce the lipophilicity of the anion, frequently coming along with less ecotoxicity.
Also for the cyano based anions, substitution of single cyano groups with other functional groups, e.g. carboxylic acids, is possible.169 However, it is questionable if such modification will lead to anions that can fulfil a similar property profile and are suitable for applications.
3. IL cations
As summarized in Chapter I, the toxicity and biodegradability of the cationic moiety has already been extensively studied. However, there is a huge knowledge gap for the biodegradation potential of IL cations owning head groups different from imidazolium and pyridinium. This was intended to be filled with the study summarized in paper No 4, 8 and 9.
Ready biodegradation was found in paper No. 4 for the ammonium based cations choline and its unsubstituted derivative (butyltrimethylammonium, N1114), whereas the methoxy-choline and triethylmethylammonium cation was not degraded at all. Contrary to this, in paper No. 8, N1114 and tributylmethylammonium were not readily biodegradable. Such differences for the same substance within different biodegradation experiments were already reported by other groups.133,137 Variations in the microbial composition of the inoculum are the most plausible explanation for this. However, this might lead to high
95
uncertainties and false negative results when evaluating the studies. OECD standards seem not to be enough to guarantee reproducible results here and need to be further improved.
The usage of a homogeneous and commercially available mixture of a freeze-dried microorganism sounds ideal, but was not successful in previous studies.132 Techniques available to identify the microbial composition, e.g. gene sequence analysis, is another option, although this is usually very time-consuming and cost-intensive.
The toxicity studies of the investigated ammonium based cations underlined the biocompatibility of choline. In all test systems, ILs containing this cation showed no adverse effects up to 100 mg L-1. The other ammonium based cations usually had similar ecotoxicological properties. Solely D. magna was affected in the presence of ILs with methoxy-choline and N1114 cations. The choice of compounds originating from natural sources is therefore a good opportunity to design sustainable products. However, their hydrophilic character limits their application potential.
In paper No. 9, it could be shown that some of the investigated cations are readily or inherently biodegradable. Above all, the pyrrolidinium-based IL cations showed enhanced biodegradation potential compared to imidazolium. Here, the compounds with shorter side chains (C=4) were fully mineralized in a prolonged test duration. In combination with the toxicological properties (similar to imidazolium based ILs), this type of head group should be intensively considered when designing sustainable ILs. The other way around is the behaviour of ILs containing a morpholinium core. Here, the biodegradation is similar to imidazolium based cations. In both cations, only the side chain is degraded, whereas the ring structure remains intact. However, it is already known that morpholinium ILs gave lower toxicities and should therefore be preferred towards imidazolium or piperidinium.
Diionic cations were shown in paper No. 5 to be, in some cases, significantly less toxic than monoionic cations with short side chains. The structure-activity-relationships known for monocations, meaning a higher toxicity for longer alkyl chains, could also be found for the dications. However, differences between the test systems were observed in the sensitivity for prolonged side or linkage chains of the dication. The biodegradation of the tested dicationic ILs was poor: none of them were primarily degraded, even if linked by long and functionalised chains. Here, the ring structure might impede a bacterial attack. The idea for
96
further design should therefore be the prolongation or functionalisation of the side chains.
Combining this with the usage of pyrrolidinium rings, found to be degraded in monocationic ILs, could enhance the biodegradation potential of dicationic ILs.
97